PROJECT SUMMARY/ABSTRACT For individuals with opioid use disorder (OUD) the propensity for relapse is high when experiencing craving, independent of a desire to stay sober. Historically, tools to fight opioid addiction have been limited and retrospective. By the time a traditional intervention occurs, they are often already reusing. Addiction relapses too often lead to a costly downward spiral - committing crimes, getting rearrested, being hospitalized, overdosing, and/or dying. Recent advances in wearable sensors, smartphones and artificial intelligence have created an unprecedented opportunity to produce positive health outcomes by predicting and preventing OUD relapses and overdoses. With sensing and associated algorithms, it will be possible to detect relapse risk states and intervene in advance - before a relapse actually occurs. The first step in this new era of proactive opioid relapse prediction is to identify and measure digital biomarkers associated with reinitiating drug use and relapse and implement a predictive model with minimal false positive alerts to achieve just-in-time intervention. This will begin with the collection of quantitative and qualitative data on patients and healthcare professionals to identify desired features and functions. Behaivior LLC will develop a relapse prevention platform that will consider a combination of physiological sensor data from the patient?s wearable devices in combination with smartphone usage data, location data and ? importantly ? reports from people in his/her support network to identify and detect relapse triggers in real-time. SA-1?: Demonstrate what makes a wrist wearable device acceptable and usable for people with OUD. SA-2.A?: Identify positive correlations between physiological, location and smartphone data, and self-reported opioid cravings: ?Results will be used to create a predictive model that can operate in real-time. SA-2.B?: Demonstrate feasibility of using support-network-reported data to predict opioid cravings. SA-3?: Create a predictive model to anticipate opioid relapse as a step towards just-in-time intervention. SA-4?: Demonstrate feasibility of predictive model by obtaining a sufficiently low false positive rate. A successful outcome of this project will be a wearables platform capable of predicting and preventing opioid addiction relapses, stemming the tide of the opioid epidemic. The platform will be ready for regulatory approvals and integration into existing electronic health records and monitoring systems. To achieve behavior modification, Behaivior will implement and commercialize a wearables platform and suite of effective interventions utilizing individualized, active machine learning, and continued iteration of next generation wearables and supporting mobile applications. ?A? robust platform prototype capable of producing measurable change will be tested. Further development of the most promising MVP(s) will be pursued by a Phase II SBIR proposal.